AutoML Seminar

Automated Machine Learning (short AutoML) progressively automates different parts of the machine learning workflow. There is a body of work that provides efficient methods to automatically optimize the hyperparameters, such as learning rates, regularization parameters, etc of machine learning methods. To go even one step further, recent work on neural architecture search designs more powerful and efficient neural network architectures in a data-driven way.

However, with the rising popularity of AutoML both in academia and industry, it also becomes increasingly harder to keep up with the latest work. The AutoML seminar is a platform to learn and discuss recent work in this field. In regular sessions we organize 40 minute talks by someone from the AutoML community followed by a 20 minute discussion. All content is free and open to anyone. If you are interested and want to join, please follow the steps listed below. The AutoML Seminar is organizes as part of the ELLIS units Berlin and Freiburg and ScaDS.AI.

Next Talk

November 21st 1pm CET - Gabriela Kadlecova Title: GRAF: Performance Prediction with Neural Graph Features

Join the Talks

Subscribe to our Google group to get all the updates about the seminars. All the information about new talks and how to connect are communicated there. Also follow us on Twitter and subscribe to our Youtube. Subscribe to our calendar to not miss any talks.

Become a Speaker

If you want to become a speaker, please send an email to automlseminar - at - gmail - dot - com or contact one of the organizers. We particularly invite junior researchers to present their work. Talks should meet the following criteria:

  • Is this content interesting for an audience of AutoML experts?
  • Is the talk based on research papers published at a top tier machine learning conference (ICMLR, NeurIPS, ICLR, AutoML Conf etc.)? We are not strict about this. However, if you want to present content which has not been published yet, please send us your draft.
  • Does the talk present an open-source library that addresses the AutoML community?

Organizers

Aaron Klein
ScaDS.AI
Arber Zela
University of Freiburg
Giovanni Zappella
Amazon Web Services
Jovita Lukasik
University of Siegen
Rhea Sukthanker
University of Freiburg
Lennart Purucker
University of Freiburg
Frank Hutter
University of Freiburg
Cedric Archambeau
Amazon Web Services